Frequently Asked Questions about Parquet MCP Server

Q: What is the Parquet MCP Server? A: The Parquet MCP (Model Control Protocol) Server is a tool that provides web search and similarity search functionalities, designed to work seamlessly with Claude Desktop and other AI platforms. It enables AI applications to access and process information from the web efficiently.

Q: What is MCP? A: MCP stands for Model Control Protocol. It’s an open protocol that standardizes how applications provide context to LLMs, allowing AI models to interact with external data sources and tools.

Q: What are the main functionalities of the Parquet MCP Server? A: The server offers two main functionalities: 1) Web Search: Perform web searches and scrape results. 2) Similarity Search: Extract relevant information from previous searches.

Q: How do I install the Parquet MCP Server? A: You can install it either via Smithery using the command npx -y @smithery/cli install @DeepSpringAI/parquet_mcp_server --client claude, or by cloning the repository, setting up a virtual environment, and installing the package manually.

Q: What environment variables do I need to configure? A: You need to create a .env file with variables like EMBEDDING_URL, OLLAMA_URL, EMBEDDING_MODEL, and API keys for search services such as SEARCHAPI_API_KEY, FIRECRAWL_API_KEY, and VOYAGE_API_KEY.

Q: How do I use the server with Claude Desktop? A: Add the server configuration to your claude_desktop_config.json file, specifying the command and arguments to run the server.

Q: What are the available tools in the server? A: The server provides two main tools: Search Web (to perform a web search and scrape results) and Extract Info from Search (to extract relevant information from previous searches).

Q: What parameters are required for the ‘Search Web’ tool? A: The required parameter is queries, which is a list of search queries. An optional parameter is page_number, which defaults to 1.

Q: What parameters are required for the ‘Extract Info from Search’ tool? A: The required parameter is queries, which is a list of search queries to merge.

Q: How do I test the MCP Server? A: You can run the test suite in the src/tests directory using python src/tests/run_tests.py or run individual tests like python src/tests/test_search_web.py.

Q: What should I do if I encounter SSL verification errors? A: Make sure the SSL settings in your .env file are correctly configured.

Q: What should I do if embeddings are not being generated? A: Check that the Ollama server is running and accessible, the specified model is available, and the text column exists in your input Parquet file.

Q: What is UBOS? A: UBOS is a full-stack AI Agent Development Platform focused on bringing AI Agents to every business department. It helps orchestrate AI Agents, connect them with enterprise data, build custom AI Agents with LLM models, and create Multi-Agent Systems.

Q: How can I integrate Parquet MCP Server with UBOS? A: By integrating the Parquet MCP Server with UBOS, you can enrich AI Agent context with real-time web information, automate data gathering and analysis, and build more intelligent and context-aware applications.

Q: Where can I find example prompts for using the server? A: Example prompts are provided in the documentation for both Web Search and Extracting Info from Search functionalities.